LLM EXPLORER 59,420 MODELS INDEXED

MMedLM by Henrychur

By Henrychur · 11 downloads
HF Repo ↗ Copy URL

MMedLM is an open-source language model by Henrychur. Features: 7.3b LLM, VRAM: 29.3GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.11.

  Arxiv:2402.13963   Custom code   Dataset:henrychur/mmedc   En   Es   Feature-extraction   Fr   Internlm   Ja   Medical   Pytorch   Region:us   Ru   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/Henrychur/MMedLM 

MMedLM Parameters and Internals

Model Type 
multilingual, medical domain
Additional Notes 
Foundation model without instruction fine-tuning.
Supported Languages 
en (English), zh (Chinese), ja (Japanese), fr (French), ru (Russian), es (Spanish)
Training Details 
Data Sources:
MMedC
Data Volume:
25.5B tokens
Methodology:
Further pretraining on MMedC
Context Length:
2048
Model Architecture:
Auto-regressive
Input Output 
Input Format:
Text
Accepted Modalities:
text
Output Format:
Text
Performance Tips:
Use transformers version 4.28.1 to avoid errors.
Release Notes 
Version:
MMedLM
Date:
2023.2.20
Notes:
Initial release of MMedLM.
Version:
MMedBench
Date:
2023.2.20
Notes:
Release of MMedBench, a multilingual medical multi-choice question-answering benchmark.
LLM NameMMedLM
Repository 🤗https://huggingface.co/Henrychur/MMedLM 
Model Size7.3b
Required VRAM29.3 GB
Updated2026-08-01
MaintainerHenrychur
Model Typeinternlm
Model Files  10.0 GB: 1-of-3   10.0 GB: 2-of-3   9.3 GB: 3-of-3   10.0 GB: 1-of-3   10.0 GB: 2-of-3   9.3 GB: 3-of-3
Supported Languagesen zh ja fr ru es
Model ArchitectureInternLMForCausalLM
Licenseapache-2.0
Context Length2048
Model Max Length2048
Transformers Version4.28.1
Is Biased1
Tokenizer ClassInternLMTokenizer
Padding Token</s>
Vocabulary Size103168
Torch Data Typefloat32